Research Scientist & Employee #3
Smartcuts is building a new category of productivity software that gives anyone the ability to automate tedious work in minutes. You record your actions just as you would a screencast, and our browser extension turns that recording into an automation you can use and share with others. Here's a 1 minute video showing our product in action.
No product today can claim to be the go-to tool that you reach for when you want to automate anything. Our goal is to be that go-to automation tool that everyone uses.
Our platform will become a universe of discoverable and easy-to-use automations, created and maintained by a community of passionate users. Over time, as a second order effect, we'll have created a crowdsourced engine that constantly maps out the workflows that take place within and between web apps to get work done.
We're currently in private beta and already have paying customers, from individual consumers to well known, publicly traded technology companies.
Smartcuts is founded by the co-founder of Lever and is backed by investors including Y Combinator, SignalFire, HOF Capital, and founders of Firebase, FitBit, HubSpot, Lyft, and Segment. We are a distributed team located across New York City, Shanghai, and the UK. If you are a maker who values craftsmanship and design, we'd love to talk with you.
We are building a new kind of automation product that will be the first of its kind -- a product that gives everyday, non-technical users superpowers to automate any repetitive work, on any website, without code (1 minute video). To do this, we have to solve challenging engineering problems.
You're someone who loves discovering and inventing new things, and you love working with a lot of autonomy on small teams. You will be responsible for our core algorithm. That means that you will be working on novel approaches for automatically extracting structured data from web pages as well as anticipating user intent. We are interested in an ensemble approach that draws from past research, machine learning, computer vision, and NLP. Bonus points if you have experience with web mining, web APIs, working with Chrome DevTools Protocol, or have an impressive mastery over the DOM and its nuances.
Our core algorithm is written in a dialect of Ocaml called ReasonML. Our choices were built on a preference for strongly typed, functional languages and for serverless architectures -- choices that allow us to focus on the necessary, not the incidental complexity of the problem space. As the lead research scientist, you can choose the language and technology stack that best gets the job done.
We use ReasonML, ReScript, Next.js, Vercel, GraphQL, and the serverless db Fauna. Our choices come from a preference for strongly typed, functional languages and for serverless architectures -- choices that allow us to focus on the necessary, not the incidental, complexity of the problem space.
Some of the interesting engineering and design challenges in this problem space?
- How do you design a product that doesn't expose engineering concepts to users?
- How can you design an algorithm that can reliably locate elements on a page, without asking the user to specify css selectors?
- How can you leverage AI to automatically detect and extract structured records from a web page?